Jerking in motion on same points of trajectory.

[shad]

Hello Balazs!
My Customer noticed that system has a big jerking in some points of the trajectory. Problem appears on some joints where arcs connects with linear segments and where linear segments connects each other.Looks like deceleration, stop and then acceleration issues. Manipulating with settings of trajectory planner like linear error and corner error doesn't help at all. I am attached program file (pdf file just change extension), printscreens of the UCCNC settings and link to the video.
Jerking was on all points were arcs connects with linear segments, for example on this places:
N2800-N2810
N2840-N2850
N7630-N7700
N21430-N21460
N21500-N21570
System: UCCNC 1.29, UC400ETH.

https://www.dropbox.com/s/ni46vb88n3v33 ... 1.mp4?dl=0

Do you have any thought?

[shad]

For example this is the photo of the cut. I am very sure that this is not THC issue.
Torch really stop on the XY plane on this points of the trajectory.

[cncdrive]

Hi Andrew
Unfortunately I can't open your testvector document, the file seems to be corrupted or something.
Could you please upload a short piece of the g-code where there is issue?
And also it would be great to have the profile file to see the CV parameters.
Without seeing the code I guess that the connection angle is larger than the CV set angle so that point makes the CV planner to slowdown or the linear unify or addition length is too short, so the planner will slowdown on that segment, but it would be the best if I can see that part of the g-code and the profile then I could check it and will have more ideas...

[shad]

Thank you for fast answer.
Yes, problem issue when there are a lot of short vectors. This is a problem for planner?
I sent all to the your email.
This is a short cut from program for example:

Code: Select all

N7630 G02 X185.3580 Y169.1951 I-0.6328 J-0.4895
N7640 G01 X186.4457 Y167.4102
N7650 G01 X186.4533 Y167.3974
N7660 G01 X187.5078 Y165.5926
N7670 G01 X187.5151 Y165.5797
N7680 G01 X188.5123 Y163.7977
N7690 G01 X188.5553 Y163.7405
N7700 G03 X189.1362 Y164.3633 I-0.2293 J0.7961

[cncdrive]

Hi Andrew, got your e-mail, thanks.

[ipmayerav]

Hello everyone? I'm the Andrew's customer who has jerking problem. You could download tap file trough the following link
https://yadi.sk/d/zr49WuMk37Xtzf

While I was testing your conroller UC400ETH I have noticed one more thing that seems to be a problem for me.
When soft limits are set there is some kind of bug with them.
When you are moving fast in direction of soft limit with only one axis it's all fine - axis stops normaly when it achieves soft limit.
If you are also moving another axis too, wich is not hitting soft limit - it also stops instantly loosing steps, when the first axis hit soft limit. I think that it is not normal situation and it should be handled differently by controller.

[A_Camera]

While I was testing your conroller UC400ETH I have noticed one more thing that seems to be a problem for me.
When soft limits are set there is some kind of bug with them.
When you are moving fast in direction of soft limit with only one axis it's all fine - axis stops normaly when it achieves soft limit.
If you are also moving another axis too, wich is not hitting soft limit - it also stops instantly loosing steps, when the first axis hit soft limit. I think that it is not normal situation and it should be handled differently by controller.

Did you try to install version 1.2031? Maybe the problem is fixed in that one since some errors were fixed which concern soft limits.

[ipmayerav]

While I was testing your conroller UC400ETH I have noticed one more thing that seems to be a problem for me.
When soft limits are set there is some kind of bug with them.
When you are moving fast in direction of soft limit with only one axis it's all fine - axis stops normaly when it achieves soft limit.
If you are also moving another axis too, wich is not hitting soft limit - it also stops instantly loosing steps, when the first axis hit soft limit. I think that it is not normal situation and it should be handled differently by controller.

Did you try to install version 1.2031? Maybe the problem is fixed in that one since some errors were fixed which concern soft limits.

Yes i did. In this version the problem remains. Developers said that they found the reason of tihis issue ad will fix it in the next versions.

[CT63]

I'm using a ETH400 and experienced the same issue losing steps randomly when the x & y axis is moving running gcode. I seems to only be the X axis.

[cncdrive]

Yes i did. In this version the problem remains. Developers said that they found the reason of tihis issue ad will fix it in the next versions.

No, this is not exactly correct.
The issue is the small around 0.2 units size lines in the g-code between the large arcs.
My collegue will continoue to investigate the issue next week if these lines should cause the planner to deccelerate or not, so far I do not know if it has to do that or not.

I'm using a ETH400 and experienced the same issue losing steps randomly when the x & y axis is moving running gcode. I seems to only be the X axis.

No, this can't make an axis to loose steps, since the decceleration is normal, happens with the value you setup for the axis, so this is no a harmful problem at all.
What happens is simply that the CV planner finds that miniature line segment to be not optimisable out in the CV planning, so it deccelerates down and accelerates up again with the normal setup acceleration/decceleration values.
The OPs thread name is missleading and is not correct, since the issue is not a jerk (change in acceleration) in the motion, but a normal decceleration...